source: branches/polybatch/tutorials/tutorial_5.rst@ 2419

Last change on this file since 2419 was 1636, checked in by Malte Marquarding, 15 years ago

Added sphinx project for ASAP tutorials

File size: 3.1 KB

Tutorial #5 - ASAP advanced

?
.. sectionauthor:: Maxim Voronkov

Main goal: To understand how to handle the data manually by building a quotient without using mx quotient

Files

  • 2009-04-02 2240 MMB-MX-11.9-0.13333.rpf data file
  • Script 1: tutorial.py general reduction script
  • Script 2: systemquotient.py script doing built-in quotient (to be copied and modified)
  • Script 3: userquotient.py the result (you’re not supposed to look into itunless desperate)

Instructions

  1. Load the data from RPFITS file
  2. Play with various methods to get access to the data
  3. Run the reduction script using the built-in quotient
  4. Copy the template script doing the quotient
  5. Modify makeQuotient method to replicate the functionality of the built-in mx quotient

Help: Use help method in asap to get information about parameters, etc

%run tutorial.py

Copy systemquotient.py to a new name, change the import statement at the top of tutorial.py to load your file instead of the systemquotient.py.

from myquotient import *

Now your version of makeQuotient will be called from tutorial.py. Use the following to load the data into a scan table called sc

sc=scantable("2009-04-02_2240_MMB-MX-11.9-0.13333.rpf")

Inspect the content.

sc.summary()

The file contains 7 scans. Each scan corresponds to observations of a source with a different beam of a 7-beam receiver. We want to use the median spectrum from all other scans as a reference when constructing the quotient using the formula

Error: Failed to load processor none
No macro or processor named 'none' found

where Toff is the system temperature measured during the reference scan, On is the signal spectrum, Off is the reference spectrum. We want to construct Result for each individual beam (out of 7 beams available) and average all these spectra together.

To do the selection

sel=sc.get_selection()
sel.set_beams(0)
sel.set_scans(0)
sel.set_polarisations(0)
sc.set_selection(sel)
selected_sc = sc.copy()
selected_sc.summary()

To average or compute the median use one of

ref.average_time(weight="median")
scans.average_time()

To merge the scans together (i.e. individual quotients for each beam) build a python list first and then use merge and average time

res=[]
for b in range(7):
    res.append(myFunctionReturningAScanTable(beam))
averaged_scantable = average_time(merge(res))

Note that merge will fail if the list has only 1 element. An if-statement may be necessary To scale the scantable with the constant factor use:

scan.scale(factor,tsys=False, insitu=True)

To add a constant use:

scan.add(constant_to_add, insitu=True)

To get the tsys use:

scan.get_tsys()

Note that some selection of data is usually necessary. Otherwise get_tsys() returns too many numbers. To divide two spectra simply divide one scan table to another:

quotient = signal / ref
Note: See TracBrowser for help on using the repository browser.